Biometrics: Steps of Iris Recognition Method

Biometrics is one of the foremost technology for flawless and error free authentication of individuals thereby offering superb security. Biometrics literal translates to measurement of humans biological characteristics that are obviously distinct and therefore used as excellent tool for identification or authentication.

Image Source - [2]

There are many different types of biometric methods available for different applications. Some of them are as follows:- facial expression identification, iris recognition, fingerprint pattern recognition, voice recognition and the most common & the oldest method - signature identification. Table below shows the list of various biometric methods with level of security each one of them offers.

Comparison of Biometric Methods

As we know quite well that every individual in our country is provided with unique identification by Unique Identification Authority of India (UIDAI) through 'Aadhaar'. Adhaar as we know uses biometric technique including fingerprint and iris scan in order to establish identity of all citizen of our country. Iris scan as we have seen in the table above has advantage over fingerprint recognition as former offers more security. In this article, various steps that are involved in iris scan are discussed briefly.

A biometric identification of iris uses mathematical pattern-recognition technique that is applied to captured images of eyes of individual. The patterns used in this technique are complex patterns. Figure below shows the block diagram of iris recognition system with each stage explained as follows:-

Block Diagram - Iris Recognition System

• Image acquisition: Iris images of individual are captured using very high resolution camera or sensor that provides good sharpness and resolution. For enhancing the quality of captured images preprocessing techniques could be applied. These preprocessing techniques are histogram equalization, noise filtering etc.
• Iris Segmentation: In this step approximation of iris region is done using two circles. For detection of eyelids the derivatives in the parallel direction are applied. For detection of outer rounded boundary of iris, the derivatives in the vertical direction are applied. Circular Hough transform is performed for segmentation and for locating the iris boundary, eyelids influence is diminished by taking only the vertical gradients.



Iris Segmentation

• Normalization: There could be certain variations or noisy data in the images of the eyes due to change in size of pupil, variable iris etc. during capturing of images. So transformation of iris area is done so that it has dimensions that are firm and unchangeable for comparisons.
• Feature Encoding: The most discriminating and unique lineaments of iris are extracted in this phase.
• Feature Matching: Next step is to match the encoded feature with the images present in iris template database. Based on this match the decision is taken.

Iris of every person is unique, it never changes during human lifetime and can never be cosseted. Therefore Iris scan could be a good biometric identification technique as it is difficult to forge and is highly accurate. And even more so, today with the advancement in electronics technology the cost of hardware and processors are coming down. So it is not very difficult or costly to replace existing RFID authentication method, which is most popular, with Iris recognition method.

By: Ms. Latika Kakkar - Assistant Professor, CSE, Chitkara University, H.P.

References

1. Wikipedia.org
2. Iritech.com